Header bidding is credited with helping drive up revenues for publishers by letting more demand sources compete simultaneously, but it has been slower to take off for video advertising says Mark Giblin, vice president BD, programmatic demand at Unruly. In this piece, Giblin looks back at the history of header bidding to explain why it’s had an uneasy transition to video, and outlines how the problems with video header bidding are being solved.
Unruly’s global programmatic lead Paul Gubbins will be on stage at New Video Frontiers in London on October 16th discussing what the digital ad ecosystem can do to cope with restrictions on cookie-based targeting.
Header bidding for video has come a long way in recent years and is becoming the dominant mechanism for publishers looking to monetise their video inventory. The success of display header bidding is well documented, and the industry is now in position for video to follow suit.
However, while header bidding in the display market has flourished, video adoption has been a lot slower, as publishers struggle to get to grips with some of the challenges monetising their video inventory presents.
But what has prevented header bidding from being as widely adopted in video? And how are these challenges being overcome?
Replicating Display’s Success
First of all, let’s look at why header bidding has become so popular.
Header bidding first emerged as a way of allowing all demand sources to compete on an equal footing, with SSPs receiving the ad call before it was sent to a publisher’s ad server.
Previously, publishers had used a waterfall set-up, which ordered SSPs based on their historical pricing data. But in an ecosystem based on impressions, this was incredibly inefficient, as it left demand partners willing to pay more for an individual impression out in the cold if they were further down the waterfall.
To add to this inefficiency, a feature in the Google-owned ad server (formerly DFP) also allowed the Google-owned ad exchange (formerly AdX) to have the last look at a line items’ CPM value and cherry pick impressions.
Header bidding changed all of that, with every impression receiving the maximum bid for that specific impression, opening up more opportunities for advertisers and ensuring publishers made more money from their inventory. While AdX still had a last-look advantage, at the very least it had to work harder to win the bid, resulting in higher CPMs for publishers.
This ability to maximise yield has led to header bidding strategies being adopted at a rapid rate, and it’s now unquestionably the standard set-up for bidding on display inventory.
This has been aided by many advances in header bidding, such as the move from the browser (client side) to the server (server side), reducing latency risk and allowing publishers to add a greater number of partners.
Unsurprisingly, having seen the impact header bidding has had on their display revenues, publishers are also keen to do the same with video. Video supply is scarcer, so it’s even more important to maximise the number of bids for every impression. But despite this, video header bidding has lagged behind that of display, and is only now beginning to gain the same levels of adoption.
Exclusive Partnerships and Technical Hurdles
The reasons for this stem from the key differences in the video and display markets.
The set-up, at least in theory, is the same: SSPs receive the ad call before the ad server and publishers receive the highest available bid for every impression. But video presents a number of unique technical and commercial challenges.
For example, publishers tend to work with specialist video partners through minimum revenue guarantees (MRGs) and exclusive partnerships. This removes the benefits of header bidding as there is no impression-level competition. This used to be the case in display as well, and all signs point to video following the same route as publishers realise the benefits of opening up competition and increasing bid density.
There are also technical hurdles. The most obvious being there’s no ‘header’ in video, so the header bidding logic had to be included inside the player (the death of Flash players has helped here).
Furthermore, any risks around display are magnified with video: the CPMs are higher; the content and creative is more expensive; the file sizes are larger etc.
Latency is also a huge concern. Wrappers such as prebid.js are continuing to evolve, with features such as improved safety timeout settings, while publishers are becoming more skilled at testing and removing partners that don’t meet their requirements.
The fact is video advertising is simply more complex than display advertising. There’s a wide variety of video players and an ever-evolving set of standards that need to be navigated when running video.
It’s meant many publishers have not had the resources needed to manage a video header bidding set-up.
But that’s changing. Prebid video has come a long way in the past two years, thanks to improvements in video integration and the introduction of new demand partners. This has lifted some of the burden for publisher dev teams by requiring less custom work, and as a result the adoption of video header bidding has been accelerating.
And it’s only going to get faster. As many advertisers ramp up their investment in video over the coming months—with some predicting it will account for as much as 80 percent of all online marketing content by the end of 2019—the timing could not be better.